摘要：The risk of lost circulation is a complex problem that cannot be ignored during drilling operations, and accurate risk assessment is crucial for preventing and controlling lost circulation events. In this study, we establish a multidimensional lost circulation quantitative risk assessment model based on ensemble machine learning, comprehensively considering three dimensions—formation risk, drilling operation risk, and fluid circulation risk. It can effectively capture and quantify the complex interactive relationship between different risk factors, and the accuracy and efficiency of lost circulation risk assessment can be improved when ensemble machine learning algorithms determine different dimensional risk weights. The results of example verification show that when the threshold of the lost circulation risk index is set to 0.55, in the set of 442 samples drilled in a certain block, the assessment accuracy is 85.02% in the samples without lost circulation and 70.21% in the samples with lost circulation. This result reflects the uncertainty of the occurrence of lost circulation events in field operations, the difference in accuracy between the two categories is approximately 15%, and this error is within an acceptable range (0.1~0.2). The independent variable parameters of each dimension of risk in the model can be adjusted according to the actual situation of different blocks, and different lost circulation index thresholds and correction factors can be set. The established model has high adaptability, which can guide lost circulation prevention and control. Copyright ? 2025 Society of Petroleum Engineers.

ISSN号：1086-055X

卷、期、页：卷30期5页2285-2295

发表日期：2025-05-01

影响因子：0.000000

期刊分区（SCI为中科院分区）：三区

收录情况：SCI（科学引文索引印刷版）,EI（工程索引）,SCIE（科学引文索引网络版）

发表期刊名称：SPE Journal

参与作者：孙金声

通讯作者：沐华艳,张伟,冯奇,王全得

第一作者：蒋官澄,贺垠博,董腾飞,杨丽丽

论文类型：期刊论文

论文概要：沐华艳,蒋官澄,张伟,孙金声,贺垠博,董腾飞,冯奇,王全得,杨丽丽，Multidimensional Lost Circulation Risk Quantification Assessment Model Based on Ensemble Machine Learning，SPE Journal，2025，卷30期5页2285-2295

论文题目：Multidimensional Lost Circulation Risk Quantification Assessment Model Based on Ensemble Machine Learning